rocSPARSE 3.1.2 for ROCm 6.1.0

Additions

• New LRB algorithm to SpMV, supporting CSR format
• rocBLAS as now an optional dependency for SDDMM algorithms
• Additional verbose output for csrgemm and bsrgemm

Optimizations

• Triangular solve with multiple rhs (SpSM, csrsm, ...) now calls SpSV, csrsv, etcetera when nrhs equals 1
• Improved user manual section Installation and Building for Linux and Windows
• Improved SpMV in CSR format on MI300

rocSPARSE 3.0.2 for ROCm 6.0.2

rocSPARSE code for ROCm 6.0.2 did not change. The library was rebuilt for the updated ROCm 6.0.2 stack.

rocSPARSE 3.0.2 for ROCm 6.0.0

Added

• Added rocsparse_inverse_permutation
• Added mixed precisions for SpVV
• Added uniform int8 precision for Gather and Scatter

Optimized

• Optimization to doti routine
• Optimization to spin-looping algorithms

Changed

• Changed rocsparse_spmv function arguments
• Changed rocsparse_xbsrmv routines function arguments
• doti, dotci, spvv, and csr2ell now require calling hipStreamSynchronize after when using host pointer mode
• Improved documentation
• Improved verbose output during argument checking on API function calls

Deprecated

• Deprecated rocsparse_spmv_ex
• Deprecated rocsparse_xbsrmv_ex routines

Removed

• Removed auto stages from spmv, spmm, spgemm, spsv, spsm, and spitsv.
• Removed rocsparse_spmm_ex routine

Fixed

• Fixed a bug in rocsparse-bench, where SpMV algorithm was not taken into account in CSR format
• Fixed the BSR/GEBSR routines bsrmv, bsrsv, bsrmm, bsrgeam, gebsrmv, gebsrmm so that block_dim==0 is considered an invalid size
• Fixed bug where passing nnz = 0 to doti or dotci did not always return a dot product of 0

rocSPARSE 2.5.4 for ROCm 5.7.1

rocSPARSE code for ROCm 5.7.1 did not change. The library was rebuilt for the updated ROCm 5.7.1 stack.

rocSPARSE 2.5.4 for ROCm 5.7.0

Added

• Added more mixed precisions for SpMV, (matrix: float, vectors: double, calculation: double) and (matrix: rocsparse_float_complex, vectors: rocsparse_double_complex, calculation: rocsparse_double_complex)
• Added support for gfx940, gfx941 and gfx942

Improved

• Fixed a bug in csrsm and bsrsm

Known Issues

In csritlu0, the algorithm rocsparse_itilu0_alg_sync_split_fusion has some accuracy issues to investigate with XNACK enabled. The fallback is rocsparse_itilu0_alg_sync_split.

rocSPARSE 2.5.2 for ROCm 5.6.1

rocSPARSE code for ROCm 5.6.1 did not change. The library was rebuilt for the updated ROCm 5.6.1 stack.

rocSPARSE 2.5.2 for ROCm 5.6.0

Improved

• Fixed a memory leak in csritsv
• Fixed a bug in csrsm and bsrsm

rocSPARSE 2.5.1 for ROCm 5.5.1

rocSPARSE code for ROCm 5.5.1 did not change. The library was rebuilt for the updated ROCm 5.5.1 stack.

rocSPARSE 2.5.1 for ROCm 5.5.0

Added

• Added bsrgemm and spgemm for BSR format
• Added bsrgeam
• Added build support for Navi32
• Added experimental hipGraph support for some rocSPARSE routines
• Added csritsv, spitsv csr iterative triangular solve
• Added mixed precisions for SpMV
• Added batched SpMM for transpose A in COO format with atomic atomic algorithm

Improved

• Optimization to csr2bsr
• Optimization to csr2csr_compress
• Optimization to csr2coo
• Optimization to gebsr2csr
• Optimization to csr2gebsr
• Fixes to documentation
• Fixes a bug in COO SpMV gridsize
• Fixes a bug in SpMM gridsize when using very large matrices

Known Issues

• In csritlu0, the algorithm rocsparse_itilu0_alg_sync_split_fusion has some accuracy issues to investigate with XNACK enabled. The fallback is rocsparse_itilu0_alg_sync_split.

rocSPARSE 2.4.0 for ROCm 5.4.4

rocSPARSE code for ROCm 5.4.4 did not change. The library was rebuilt for the updated ROCm 5.4.4 stack.